This past year we completed several clinical and brain imaging investigations of adolescents with Autism Spectrum Disorders (ASD). Our clinical studies focused on a number of ASD-related behavioral characteristics and issues. In one study we showed that although high-functioning adolescents with ASD tended to have more intense and idiosyncratic interests and hobbies than age and IQ matched control subjects, their interests were not more restricted in number, as has been commonly thought. In fact, it was the intensity with which they pursued these interests and hobbies, rather than their total number, that was most strongly related to other ASD symptoms. In a separate study we found that the speed at which subjects performed relatively simple cognitive tasks was strongly related to ASD symptoms including difficulties with social communication. This finding suggests that speed of performance may be a useful predictor of functional outcome in ASD. Finally, contrary to recent claims that symptoms of depression and anxiety are associated with higher IQ and fewer ASD symptoms, we observed an increased risk for depression/anxiety symptoms in our high-functioning children and adolescents with ASD, regardless of IQ or ASD symptoms. Our structural brain imaging studies have continued to document abnormally thin cortex, as well as increased cortical gyrification, in ASD subjects relative to matched control subjects. These findings are consistent with an ever-growing literature on cortical atypicalities in ASD. We have also continued to document the functional consequences of these atypicalities using functional MRI. In a series of studies, we found that the neural circuitry associated with perceiving and understanding social interactions showed a lack of category-specificity in the ASD subjects. Specifically, brain regions that typically respond to complex motion, such as when viewing social, relative to mechanical interactions, responded equally strong to both types of interactions in ASD subjects. One interpretation of this finding is that ASD may be characterized by deficient neural connectivity within brain regions comprising the social processing network. Our studies of functional brain connectivity supported this possibility. In collaboration with the NIMH Scientific and Statistical Computing Core, we developed an unbiased, data-driven, procedure for identifying brain regions showing abnormal patterns of functional connectivity using resting-state fMRI data. During a resting-state scan, subjects simply lie quietly in the scanner, not performing any specific task. Under these conditions, spatially distant regions of the brain can show similar patterns of slowly fluctuating neural activity. The strength of similarity, or correlation, between the fluctuations in different brain regions can be used as a measure of the strength of the connections between them. Our study identified several regions of the ASD brain that showed abnormally weak connections with other brain regions. These regions included the medial portions of prefrontal cortex, the amygdala, and anterior and posterior regions of the temporal lobes, all of which have been strongly associated with social cognition. Importantly, the strength of the connections we observed was related to social deficits of our ASD subjects as determined by behavioral ratings provided by their parents - the greater the social difficulty the weaker the functional connectivity. Resting-state fMRI data have proven to be particularly useful for comparing clinical patient groups because they are easy to obtain and are not subject to group differences in performance that can confound the interpretation of task-based neuroimaging data. However, the analysis of these data is fraught with a problems and pitfalls that have not been fully appreciated or explored. To help remedy this situation, we have published several papers that describe these difficulties in detail and offer potential solutions.